A company uses Amazon Bedrock to implement a generative AI assistant on a website. The AI assistant helps customers with product recommendations and purchasing decisions.
The company wants to measure the direct impact of the AI assistant on sales performance.
Which metric will meet these requirements?
A. The conversion rate of customers who purchase products after AI assistant interactions.
B. The number of customer interactions with the AI assistant
C. Sentiment analysis scores from customer feedback after AI assistant interactions
D. Natural language understanding accuracy rates
Show Answer
Correct Answer: A
Explanation: The metric that directly measures the AI assistant's impact on sales is the conversion rate of customers who make a purchase after interacting with the assistant. The other options measure engagement, customer perception, or technical model performance rather than business outcomes tied to revenue.
Question 215
A customer service team is developing an application to analyze customer feedback and automatically classify the feedback into different categories. The categories include product quality, customer service, and delivery experience.
Which A1 concept does this scenario present?
A. Computer vision
B. Natural language processing (NLP)
C. Recommendation systems
D. Fraud detection
Show Answer
Correct Answer: B
Explanation: The scenario involves analyzing text-based customer feedback and classifying it into categories such as product quality, customer service, and delivery experience. This is a standard natural language processing (NLP) text classification task. It does not involve image analysis (computer vision), personalized suggestions (recommendation systems), or identifying deceptive activity (fraud detection).
Question 216
An education company is building a chatbot whose target audience is teenagers. The company is training a custom large language model (LLM). The company wants the chatbot to speak in the target audience's language style by using creative spelling and shortened words.
Which metric will assess the LLM's performance?
A. F1 score
B. BERTScore
C. Recall-Oriented Understudy for Gisting Evaluation (ROUGE)
D. Bilingual Evaluation Understudy (BLEU) score
Show Answer
Correct Answer: B
Explanation: BERTScore compares generated text with reference text using contextual embeddings, capturing semantic similarity and tolerating paraphrases, slang, creative spelling, and shortened words better than exact n-gram overlap metrics. BLEU and ROUGE rely largely on surface-form overlap, and F1 is typically used for classification or token-level extraction tasks. Therefore BERTScore is the most appropriate metric for evaluating this chatbot's outputs.
Question 217
A company wants to create a chatbot that answers questions about human resources policies. The company is using a large language model (LLM) and has a large digital documentation base.
Which technique should the company use to optimize the generated responses?
A. Use Retrieval Augmented Generation (RAG).
B. Use few-shot prompting.
C. Set the temperature to 1.
D. Decrease the token size.
Show Answer
Correct Answer: A
Explanation: Retrieval Augmented Generation (RAG) is the appropriate technique when an LLM needs to answer questions using a large internal documentation base. RAG retrieves relevant HR policy documents at query time and grounds the model's responses in that content, improving accuracy and reducing hallucinations. Few-shot prompting teaches patterns but does not connect the model to a knowledge base. Setting temperature to 1 increases randomness, and decreasing token size does not optimize factual responses.
Question 218
A food service company wants to collect a dataset to predict customer food preferences. The company wants to ensure that the food preferences of all demographics are included in the data.
Which dataset characteristic does this scenario present?
A. Accuracy
B. Diversity
C. Recency bias
D. Reliability
Show Answer
Correct Answer: B
Explanation: The scenario emphasizes collecting data that includes the food preferences of all demographics so the dataset represents the full population of interest. That is a characteristic of diversity, not accuracy, recency bias, or reliability.
Question 219
HOTSPOT
-
A company is using Amazon SageMaker to develop AI models.
Select the correct SageMaker feature or resource from the following list for each step in the AI model lifecycle workflow. Each SageMaker feature or resource should be selected one time or not at all.
Show Answer
Correct Answer: Managing different versions of the model: SageMaker Model Registry
Using the current model to make predictions: SageMaker Serverless Inference
Explanation: Model Registry stores and manages model versions and approvals. Serverless Inference deploys models for on-demand predictions. SageMaker Clarify is for bias detection and explainability, not versioning or inference.
Question 220
A company has a generative AI application that uses a pre-trained foundation model (FM) on Amazon Bedrock. The company wants the FM to include more context by using company information.
Which solution meets these requirements MOST cost-effectively?
A. Use Amazon Bedrock Knowledge Bases.
B. Choose a different FM on Amazon Bedrock.
C. Use Amazon Bedrock Agents.
D. Deploy a custom model on Amazon Bedrock.
Show Answer
Correct Answer: A
Explanation: Amazon Bedrock Knowledge Bases is the most cost-effective solution because it adds company-specific context to a pre-trained foundation model using retrieval-augmented generation (RAG) without requiring retraining or deploying a custom model. Agents orchestrate workflows rather than primarily supplying knowledge, switching foundation models does not incorporate proprietary company data, and custom models are significantly more expensive and operationally complex.
Question 221
A financial company has deployed an ML model to predict customer churn. The model has been running in production for 1 week. The company wants to evaluate how accurately the model predicts churn compared to actual customer behavior.
Which metric meets these requirements?
A. Root mean squared error (RMSE)
B. Return on investment (ROI)
C. F1 score
D. Bilingual Evaluation Understudy (BLEU) score
Show Answer
Correct Answer: C
Explanation: Customer churn prediction is a binary classification task. After comparing predicted churn labels with actual customer outcomes, the F1 score is an appropriate evaluation metric because it combines precision and recall into a single measure, especially useful when the churn class is imbalanced. RMSE is for regression, ROI is a business metric rather than a model accuracy metric, and BLEU evaluates machine translation quality.
Question 222
A company is integrating AI into its employee recruitment and hiring solution. The company wants to mitigate bias risks and ensure responsible AI practices while prioritizing equitable hiring decisions.
Which core dimensions of responsible AI should the company consider? (Choose two.)
A. Fairness
B. Tolerance
C. Flexibility
D. Open source
E. Transparency
Show Answer
Correct Answer: A, E
Explanation: The core responsible AI dimensions most directly relevant to mitigating bias and supporting equitable hiring are Fairness and Transparency. Fairness focuses on reducing discriminatory outcomes and ensuring equitable treatment across groups. Transparency provides explainability and clarity about how AI-assisted hiring decisions are made, supporting accountability and trust. The other options are not recognized core responsible AI dimensions in this context.
Question 223
A company has developed custom computer vision models. The company needs a user-friendly interface for data labeling to minimize model mistakes on new real-world data.
Which AWS service, feature, or tool meets these requirements?
A. Amazon SageMaker Ground Truth
B. Amazon SageMaker Canvas
C. Amazon Bedrock playground
D. Amazon Bedrock Agents
Show Answer
Correct Answer: A
Explanation: Amazon SageMaker Ground Truth provides a user-friendly data labeling interface for datasets, especially computer vision tasks such as image classification, object detection, and segmentation. It supports human-in-the-loop labeling, automated labeling, and active learning workflows to improve model performance on new real-world data. Amazon SageMaker Canvas is for no-code model building, while Amazon Bedrock playground and Bedrock Agents are for generative AI experimentation and agent workflows, not data labeling.
$19
Get all 443 questions with detailed answers and explanations
Instant download HTML + PDF delivered the moment payment clears.
Secure Stripe checkout we never see or store your card details.
7-day refund if files are defective see our refund policy.
